Add screen-adjusted device orientation attributes to DeviceOrientationEvent interface

[richtr]

Add screenAlpha, screenBeta and screenGamma attributes to the DeviceOrientationEvent interface.

This spec addition was proposed and discussed on the W3C WebApps mailing list at http://lists.w3.org/Archives/Public/public-webapps/2014JulSep/0203.html. Mozilla have expressed support to implement this and we have a patch ready for Chromium for this also (pending addition of these attributes to the spec).

[anssiko]

Generic Sensor API defines a local coordinate system that can be either the device coordinate system or the screen coordinate system.

The Accelerometer, Gyroscope, and Magnetometer specifications that decompose the DeviceOrientationEvent and DeviceMotionEvent into low-level primitives as well as the fusion sensor Orientation Sensor use these concepts, and allow the web developer to define whether the local coordinate system is to be bound to the device coordinate system or the screen coordinate system addressing this issue.

https://w3c.github.io/accelerometer/#enumdef-accelerometerlocalcoordinatesystem
https://w3c.github.io/gyroscope/#enumdef-gyroscopelocalcoordinatesystem
https://w3c.github.io/magnetometer/#enumdef-magnetometerlocalcoordinatesystem
https://w3c.github.io/orientation-sensor/#enumdef-orientationsensorlocalcoordinatesystem

Proposal: Since the Generic Sensor-based APIs now provide a modern API surface that satisfies this requirement, close this PR and do not add this new feature to the DeviceOrientation Event spec. Any concerns, please let us know in the PR comments.

[anssiko]

Hearing no concerns, it looks like this issue can be closed.

But before closing, I'd like to get feedback whether we should add some informative text along the following lines to the DeviceOrientation Event spec:

Web developers looking to define whether the sensor readings are to be resolved in relation to the screen or the device orientation are advised to consult the Accelerometer, Gyroscope, and Magnetometer specifications that decompose the DeviceOrientationEvent and DeviceMotionEvent into low-level primitives as well as the fusion sensor Orientation Sensor. These specifications specify a local coordinate system that can be either the device coordinate system or the screen coordinate system.

[TolMera]

Informative text would be useful yes, however an example jsfiddle or something of that nature that would allow someone to see when their local device does not resolve in the expected manner would be extra useful, thought this probably falls outside the scope of what you may want to provide here. Not all browsers/devices allow resolution of orientation information in a uniform manner, so I would have myself supported the concept of a "screenAlpha et al"

[anssiko]

@TolMera, thanks for the feedback! Couple of alternatives building upon the existing work:

• Add local coordinate system tests in the cross-browser test suite (e.g. https://wpt.fyi/results/gyroscope) (source) for this feature. Looping in @Honry to check. To be able to automate this, probably need an extension to the WebDriver API too.

• Add a new test to the Sensor tester (source) tool that is built exactly for this, running functional tests.

Contributions welcome!

[Honry]

@anssiko, actually we've been working on these,

• w-p-t: Bring sensor reading update function based on Chromium mock sensor implementations, [sensors] Bring setUpdateSensorReadingFunction from sensor mojo inter… web-platform-tests/wpt#11600, which including tests for screen coordinate system. Need further review.
• Sensor tester: test screen coordinate system for accelerometer, Add accelerometer screen tests in sensor-tester intel/generic-sensor-demos#50, need further review.

Maybe @rakuco would like to help review, friendly ping. :)

[reillyeon]

Discussed at Devices and Sensors WG Face-to-face at TPAC 2018. Closing this issue as we are not adding new features that are not implemented by any vendor.